The traditional login mode based on a password or a gesture is mediocre in terms of safety. Once the password or the gesture is known to others, the others can also log in. Since the voiceprint information of a user is utilized in voiceprint verification, the safety is enhanced. In the prior art, a voiceprint identification method is used for verifying the identity of the login user, mainly including a voiceprint verification method in which a random numerical string/arbitrary text and a keyboard password are combined.
Specifically: during a user registration, a client terminal receives a first voice signal generated when the user pronounces a password and a second voice signal produced when the user pronounces a random numerical string/arbitrary text generated by a server, the client terminal transmits the received voice signals to the server, the server extracts the password from the first voice signal, stores the password and obtains the acoustic characteristics of the user based on the received second voice signal.
During a user login, the client terminal receives a third voice signal generated when the user pronounces a password and a fourth voice signal produced when the user pronounces a random numerical string/arbitrary text generated by server, the client terminal transmits the received voice signals to the server, the server obtains the acoustic characteristics based on the received fourth voice signal, compares the acoustic characteristics with the acoustic characteristics obtained during the user registration, extracts the password from the third voice signal and compares the password with the password stored during the user registration, if both comparison results are consistent, the user identity verification is successful, otherwise, the user identity verification is failed. The prior art has the following defects: during the user registration and the user login, the user has to pronounce the passwords which may be exposed, resulting in low safety.